The problem I intend to address is the function of sleep, a biological puzzle that is still unsolved. Specifically, I intend to test a comprehensive, novel hypothesis about what sleep is for, called the synaptic homeostasis hypothesis (Tononi and Cirelli, Brain Res Bull 2003, Sleep Medicine Rev., 2005). The hypothesis accounts for many facts about sleep and its regulation, from behavioral and phylogenetic evidence to electrophysiological and molecular observations, and has fundamental implications for health and disease. The hypothesis makes intriguing predictions that are relevant for both basic and clinical neuroscience, and I propose to test such predictions using several complementary approaches. If the hypothesis survives the tests, we may have for the first time a solid scientific explanation of why we need to sleep. Understanding the function of sleep is obviously important both scientifically and from the perspective of human health. Sleep is a pervasive, universal, and fundamental behavior: It occupies a third of our life, and an even larger proportion in infants;it is present in every animal species where it has been studied, from fruit flies to humans;it is tightly regulated, as indicated by the irresistible mounting of sleep pressure after prolonged wakefulness;and even partial deprivation of sleep has serious consequences on cognition, mood, and health. While all available evidence indicates that sleep is of the brain, by the brain, and for the brain, the function of sleep remains unknown despite decades of intensive research. No comprehensive theory had been advanced so far, which is why the synaptic homeostasis hypothesis, if corroborated, may constitute a much needed breakthrough. In fact, the lack of understanding of why we need to sleep is problematic not only from a scientific viewpoint, but also because of its vast implications for public health. Millions of people complain of sleep problems, from insomnia to excessive daytime sleepiness, from chronic fatigue to irritability associated with unsatisfactory sleep. Sleep problems are an important aspect of several psychiatric disorders, notably mood disorders and anxiety disorders. Finally, sleep deprivation has high social costs, from driving and work-related accidents to chronically poor performance. A large segment of the population is therefore treated routinely with drugs aimed at improving sleep, or at maintaining wakefulness in the face of sleep pressure. However, such treatments are hampered by our ignorance concerning the functions of sleep. Which sleep disturbances should be taken seriously because they reflect a functional impairment and which, if any, do not interfere significantly with the functions carried out by sleep? Which abnormalities of sleep are likely to have neurobiological consequences that can lead to psychiatric disorders such as depression? And finally, what aspect of sleep should be enhanced by pharmacological or behavioral treatments, and what indices should we consider to determine their effectiveness?